SURVIVAL AND GROWTH OF ENTEROBACTER SAKAZAKII ON PRODUCE, CONDITIONS AFFECTING BIOFILM FORMATION, AND ITS SENSITIVITY TO SANITIZERS HOIKYUNG KIM

Transcription

1 SURVIVAL AND GROWTH OF ENTEROBACTER SAKAZAKII ON PRODUCE, CONDITIONS AFFECTING BIOFILM FORMATION, AND ITS SENSITIVITY TO SANITIZERS by HOIKYUNG KIM (Under the direction of LARRY R. BEUCHAT) ABSTRACT Survival and growth of Enterobacter sakazakii on produce and conditions affecting its attachment and biofilm formation were investigated. Sensitivity of the pathogen to sanitizers and disinfectants was determined. E. sakazakii grew on fresh-cut produce and in unpasteurized fruit and vegetable juice stored at and C, and survived at least days but did not grow at C. Populations significantly decreased on whole produce stored at,, and C. Retention of viability was enhanced at refrigerator temperatures. Treatment of apples, tomatoes, and lettuce with chlorine, chlorine dioxide, and Tsunami, a peroxyacetic acid sanitizer, caused significant reductions in populations of E. sakazakii, although the extent of lethality depended on the type of produce, sanitizer concentration, and treatment time. Attachment of E. sakazakii to enteral feeding tubes and stainless steel was enhanced at C compared to C. The pathogen formed biofilm on enteral feeding tubes and stainless steel at C when immersed in infant formula broth but not in tryptic soy broth or lettuce juice broth; biofilm was not formed at C. When the surface of stainless steel was spot inoculated with infant formula or water containing E. sakazakii, dried, and exposed to a relative humidity of % at,, and C, the pathogen

2 survived for at least days. Survival was enhanced at C, compared to and C, and when formula rather than water was used as a carrier. The efficacy of sanitizers in killing E. sakazakii in dried inoculum and in biofilm on the surface of stainless steel varied, depending on the composition of the carrier used to suspend cells before drying, type of sanitizer, and treatment time. The overall order of resistance of E. sakazakii to disinfectants routinely used in hospitals, day-care centers, and food service kitchens was planktonic cells < cells spot inoculated and dried on stainless steel < cells in biofilms on stainless steel. This study provides information useful in assessing the potential of produce to serve as a vehicle for E. sakazakii infections, understanding attachment and biofilm formation on abiotic surfaces, and evaluating the effectiveness of sanitizers and disinfectants for its elimination. INDEX WORDS: Enterobacter sakazakii, Fruit, Vegetable, Apple, Cantaloupe, Strawberry, Watermelon, Cabbage, Carrot, Cucumber, Lettuce, Tomato, Juice, Lactic acid bacteria, Molds, Yeasts, Chlorine, Chlorine dioxide, Peroxyacetic acid, Attachment, Biofilm, Stainless steel, Enteral tube, Infant formula, Sanitizer, Disinfectant, Quaternary ammonium compounds, Phenolic compounds, Hydrogen peroxide

3 SURVIVAL AND GROWTH OF ENTEROBACTER SAKAZAKII ON PRODUCE, CONDITIONS AFFECTING BIOFILM FORMATION, AND ITS SENSITIVITY TO SANITIZERS by HOIKYUNG KIM B. S., Sunchon National University, Korea, M. S., University of Arkansas at Fayetteville, Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY ATHENS, GEORGIA

4 Hoikyung Kim All Rights Reserved

5 SURVIVAL AND GROWTH OF ENTEROBACTER SAKAZAKII ON PRODUCE, CONDITIONS AFFECTING BIOFILM FORMATION, AND ITS SENSITIVITY TO SANITIZERS by HOIKYUNG KIM Major Professor: Larry R. Beuchat Committee: Mark E. Berrang Mark A. Harrison Joseph F. Frank Ynes R. Ortega Electronic Version Approved: Maureen Grasso Dean of the Graduate school The University of Georgia August

6 DEDICATION First, I would like to dedicate this dissertation to my parents, Mr. Yeonsik Kim and Ms. Yeoja Park. Mom and Dad, Thank you for believing me that I made a right decision when I decided pursuing my degrees in U. S. Without your unconditional love and support, I had never even thought of beginning my study. I love you. I also would like to dedicate this dissertation to my advisor and respected professor, Dr. Larry R. Beuchat. He has been an advisor, mentor, role model to me. Dr. Beuchat, Thank you for your patience and guidance, and showing me the way a scientist should be. I will never forget it. Finally, I would dedicate this dissertation to my husband, Dr. Jee-Hoon Ryu. Without his encouragement and support, I would have never even dreamed of pursuing my Ph. D. He has always made me happy and encouraged. Obba, I love you more than I express in words and thank you for always being there for me. iv

7 ACKNOWLEDGEMENTS I would acknowledge my committee members, Dr. Mark Harrison, Dr, Mark Berrang, and Dr. Ynes Ortega. A special thank to Dr. Joseph Frank for being my committee member and his advice in my research. I also deeply appreciate to the Center for Food Safety and people in this research center. Without all the facilities and people in this center, my dissertation would not be possible. Thank you to my friends, Minjae Kang, Jungsun Kim, Eunyoung Hong, in Korea. They have always made me smile and still there for me as my best friends. I am also grateful to Insook Son and Byunghee Kim for being my colleagues and Korean friends in Athens, Georgia. A special thank to Dr. Inhyu Bae in Sunchon National University, Korea, for his encouragement. I also thank my parents in law, Mr. Wonku Ryu and Ms. Choonbock Lee, for their encouragement and consideration. I am grateful to my brother, Haewan Kim, for being my only brother and sometime a friend to me. Finally, I appreciate my folks, Audrey Kreske, Jennifer Simmons, Kim Hortz, Barbara Adler, David Harrison, David Mann, Li-chun Lin, Kristen bray, Patrick Bray, Evelyn Dixon, Bobby Goss for their supportive works. A special thank must go to Joshua Gurtler. He has always been a good friend and colleague to me. Without all of them, my dissertation would not be made in this short period of time. v

8 TABLE OF CONTENTS Page ACKNOWLEDGEMENTS...v LIST OF TABLES... vii LIST OF FIGURES... ix CHAPTER INTRODUCTION AND LITERATURE REVIEW... SURVIVAL AND GROWTH OF ENTEROBACTER SAKAZAKII ON FRESH- CUT FRUITS AND VEGETABLES AND IN UNPASTEURIZED JUICES AS AFFECTED BY STORAGE TEMPERATURE... SURVIVAL OF ENTEROBACTER SAKAZAKII ON FRESH PRODUCE AS AFECTED BY TEMPERATURE, AND EFFECTIVENESS OF SANITIZERS FOR ITS ELIMINATION... ATTACHMENT AND BIOFILM FORMATION BY ENTEROBACTER SAKAZAKII ON STAINLESS STEEL AND ENTERAL FEEDING TUBES... EFFECTIVENESS OF DISINFECTANTS IN KILLING ENTEROBACTER SAKAZAKII IN SUSPENSION, DRIED ON THE SURFACE OF STAINLESS STEEL, AND IN BIOFILM... SUMMARY AND CONCLUSIONS... vi

9 LIST OF TABLES Page Table -: Differences of biochemical reactions between E. cloacae and E. sakazakii... Table -: Partial list of outbreaks and cases of E. sakazakii infections documented since... Table -: ph values of fresh-cut produce surfaces as affected by storage time and temperature... Table -: Populations of E. sakazakii recovered from fruit and vegetable juice stored at C for days... Table -: ph values of fruit and vegetable juice as affected by storage time and temperature... Table -: Brix values of fruit and vegetable juice as affected by storage time and temperature... Table -: Population of E. sakazakii recovered from produce stored at,, and C for up to days... Table -: Populations of E. sakazakii recovered from apples treated with water or sanitizers for min... Table -: Populations of E. sakazakii recovered from apples treated with water or sanitizers for min... Table -: Populations of E. sakazakii recovered from tomatoes treated with water or sanitizers for min... vii

10 Table -: Populations of E. sakazakii recovered from tomatoes treated with water or sanitizers for min... Table -: Populations of E. sakazakii recovered from lettuce treated with water or sanitizers for min... Table -: Populations of E. sakazakii recovered from lettuce treated with water or sanitizers for min... Table -: Populations of E. sakazakii in suspension and attached to surfaces of enteral feeding tubes immersed in TSB, IFB, and LJB at and C for h... Table -: Populations of E. sakazakii in suspension and attached to surfaces of stainless steel coupons immersed in TSB, IFB, and LJB at and C for h... Table -: Disinfectants evaluated for lethality to E. sakazakii... Table -: Survival of planktonic cells of E. sakazakii strain as affected by treatment with disinfectants... Table -: Survival of planktonic cells of E. sakazakii strain as affected by treatment with disinfectants... Table -: Survival of E. sakazakii strains and spot inoculated and dried on the surface of stainless steel coupons as affected by the type of carrier (water or infant formula) in which cells were suspended... Table -: Survival of E. sakazakii strains and in biofilm formed on the surface of stainless steel coupons immersed in infant formula at C for or days as affected by treatment with disinfectants... viii

11 LIST OF FIGURES Page Figure -: Procedure of E. sakazakii isolation from powdered infant formula... Figure -: Populations of E. sakazakii recovered from fresh-cut apples, cantaloupe, strawberries, watermelon, cabbage, carrots, cucumbers, lettuce, and tomatoes stored at, or C for up to days... Figure -: Populations of E. sakazakii recovered from apple, cantaloupe, strawberry, watermelon, cabbage, carrot, cucumber, lettuce, and tomato juice stored at, or C for days... Figure -: Populations of mesophilic aerobic bacteria (total counts) recovered from apples, cantaloupe, strawberries, watermelon, cabbage, carrots, cucumbers, lettuce, and tomato juice inoculated with E. sakazakii and stored at, or C for days... Figure -: Populations of lactic acid bacteria recovered from apples, cantaloupe, strawberries, watermelon, cabbage, carrots, cucumbers, lettuce, and tomato juice inoculated with E. sakazakii and stored at, or C for days... Figure -: Populations of molds or yeasts recovered from apples, cantaloupe, strawberries, watermelon, cabbage, carrots, cucumbers, lettuce, and tomato juice inoculated with E. sakazakii and stored at, or C for days... Figure -: The ph of surface tissue and internal tissue of apples, cantaloupes, strawberries, lettuce, and tomatoes inoculated with E. sakazakii and stored at C, C, and C for up to days... ix

12 Figure -: Populations of E. sakazakii strains,,,,, and grown in TSB, IFB, and LJB at C and C for up to days... Figure -: Populations of E. sakazakii strains and attached to the surface or in biofilms formed on the surface of enteral feeding tubes immersed in TSB, IFB, and LJB and in broths in which tubes were immersed and incubated at C for up to days... Figure -: Populations of E. sakazakii strains and attached to the surface or in biofilms formed on the surface of stainless steel coupons immersed in TSB, IFB, and LJB and in broths in which coupons were immersed and incubated at C for up to days... Figure -: Populations of E. sakazakii strains and attached to the surface or in biofilms formed on the surface of enteral feeding tubes immersed in TSB, IFB, and LJB and in broths in which tubes were immersed and incubated at C for up to days... Figure -: Populations of E. sakazakii strains and attached to the surface or in biofilms formed on the surface of stainless steel coupons immersed in TSB, IFB, and LJB and in broths in which coupons were immersed and incubated at C for up to days. Bars indicate standard deviations... Figure -: Populations of E. sakazakii strains and suspended in sterile distilled water and reconstituted infant formula, inoculated and dried on the surface of stainless steel, and incubated at,, and C for up to days... x

13 CHAPTER INTRODUCTION AND LITERATURE REVIEW

14 INTRODUCTION Enterobacter sakazakii infections rarely occur, compared to infections caused by several other foodborne pathogens; however, the bacterium can be life-threatening to neonates, infants, and immunocompromised adults. E. sakazakii has been categorized by the International Commission for Microbiological Specifications for Foods (ICMSF, ) as a Severe hazard for restricted populations, life-threatening, or substantial chronic sequence or long duration in a ranking of foodborne pathogens and toxins into hazard groups. The most common food vehicle associated with E. sakazakii infections is powdered infant formula. In surveys done to determine the presence of E. sakazakii in powdered infant formula, the organism was detected in. -.% of the products tested (Muytjens et al., ; Iversen and Forsythe, ). However, the organism has also been isolated from other foods, including lettuce (Soriano et al., ), tomatoes (Jung and Park, ), alfalfa sprouts (Cruz et al., ), other vegetables, cheese, minced beef, and sausage meat (Leclercq et al., ), and in household and food factory environments (Kandhai et al., ). The presence of E. sakazakii in produce raises possibilities of an increased risk of infection in elderly or immunocompromised people. Foodborne disease outbreaks associated with consumption of fresh produce have been caused by several genera in family Enterobacteriaceae. Therefore, the presence of E. sakazakii on produce and its ability to cause infections in elderly, immunocompromised people raises the need to know more about its behavior on produce. In addition to its potential lethality and apparent ubiquity, E. sakazakii has been reported to form biofilms (Iversen et al., ; Lehner et al., ). E. sakazakii has been detected on utensils used for infant formula preparation in hospitals (Simmons et al. ; Clark et al., ;

15 Noriega et al., ; Bar-Oz et al., ). Attachment and biofilm by E. sakazakii on food contact surfaces may increase the risk of E. sakazakii infections. Various sanitizers and disinfectants have been effectively used to kill or eliminate foodborne pathogens that may be contaminants in produce processing environments and in food preparation areas in hospitals. However, the efficacy of these sanitizers and disinfectants in killing E. sakazakii has not been reported. This information would be useful when developing strategies to reduce the risk of E. sakazakii infections. ENTEROBACTER SAKAZAKII General characteristics E. sakazakii was initially referred to as yellow pigmented Enterobacter cloacae and was first reported to cause an outbreak of infections in (Urmenyi and Franklin, ). Later, it was proposed that E. sakazakii be distinguished from E. cloacae as a separate species based on differences in DNA relatedness, pigment production, biotyping, and antibiotic susceptibility (Farmer et al., ; Izard et al., ). The two bacteria have many similar biochemical reactions. E. sakazakii, however, produces a yellow pigment and does not ferment D-sorbitol (Table -). E. sakazakii is a gram-negative rod, motile, and non-sporulating bacterium known to cause meningitis (Gallagher and Ball, ; Burdette and Santos, ), sepsis (Simmons et al., ), bacteremia (Noriega et al., ), and necrotizing enterocolitis (Van Acker et al., ) in preterm neonates and immunocompromised adults (Jimenez, et al., ; Pribyl et al., ; Hawkins, et al., ; Emery and Weymouth, ; Lai, ).

16 Table -. Differences of biochemical reactions between E. cloacae and E. sakazakii a (Modified from Farmer and Kelly, ; Nazarowec-White and Farber, a) Biochemical reaction E. sakazakii E. cloacae Yellow pigmentation + Fermentation of: sucrose + + dulcitol ( ) adonitol ( ) D-sorbitol + raffinose + + -methyl-d-glucoside + (+) D-arabitol ( ) Lysine decarboxylase Arginine dihydrolase + + Ornithine decarboxylase + + KCN, growth in + + a +: - % positive; (+): - % positive; ( ): - % positive; : - % positive. A mortality rate of - % has been reported (Urmenyi and Franklin, ; Joker et al., ; Adamson and Rogers, ; Kleiman et al., ; Muytjens et al., ; Willis and Robinson, ; Nazarowec-White and Farber, a). The oral infectious dose has been approximated to range from (Iversen and Forsythe, ) to CFU (Pagotto et al., ). However, pathogenesis and virulence factors of E. sakazakii have not been fully characterized. The yellow pigmentation of E. sakazakii is a unique characteristic distinguished from other Enterobacter species. The pigment is produced in greater quantities at C than at C

17 (Nazarowec-White and Farber, b). Growth characteristics of E. sakazakii in infant formula and laboratory media have been determined. E. sakazakii is capable of growing at temperatures as low as. C in brain heart infusion (BHI) broth and as high as C in infant formula milk and BHI broth (Nazarowec-White and Farber, b; Breeuwer et al., ; Iversen et al., ). When E. sakazakii was grown in reconstituted infant formula at C, the lag time and generation time were. h and. h, respectively (Nazarowec-White and Farber, b). The optimum growth temperature of six strains of E. sakazakii in whitely impedance broth, BHI broth, infant formula milk, and tryptic soy broth ranged from to C (Iversen et al., ). An environmental reservoir of E. sakazakii has not been defined. However, several possible reservoirs exist. E. sakazakii was isolated from - % of test samples of environments in households and food factories (milk powder, chocolate, potato, and pasta) (Kandhai et al., ). Other studies showed the presence of E. sakazakii in Mexican fruit flies Anastrepha ludens (Kuzina et al., ) and stable flies Stomoxys calcitrans (Hamilton et al., ). This indicates that the bacterium is wide spread in the environment. Iversen and Forsythe () demonstrated that soil, water, and vegetables may be the principal sources of E. sakazakii contamination because the organism is not a part of the normal animal and human gut biota. A method for isolating and enumerating E. sakazakii in dehydrated powdered infant formula has been developed by the U. S. FDA (). The method includes pre-enriching samples in distilled water, enriching in Enterobacteriaceae enrichment (EE) broth, surface plating on violet red bile glucose (VRBG) agar, and confirming the presumptive colonies with API E bioassay kit (Figure -).

18 Add powdered infant formula into distilled water (:) Incubate overnight at C Add ml of pre-enriched sample into ml of EE broth Incubate overnight at C Surface plate on VRBG agar (direct spreading or streaking) Incubate overnight at C Streak five presumptive colonies from the VRBG agar on TSA Incubate - h at C Confirm yellow pigmented colonies with API E Figure -. Procedure of E. sakazakii isolation from powdered infant formula (U. S. FDA, ) Epidemiology At least cases of E. sakazakii infections, resulting in deaths of neonates, infants, and young children have been documented (Iversen and Forsythe, ), while at least cases have been reported among adults (Jimenez and Gimenez, ; Pribyl et al., ; Hawkins et al., ; Emery and Weymouth, ; Lai, ; Dennison and Morris, ). Table - lists outbreaks and cases of E. sakazakii infections that have been documented since. The first two reported cases of E. sakazakii infections causing neonatal meningitis occurred in in England, but the causative bacterium was considered to be E. cloacae

19 Table -. Partial list of outbreaks and cases of E. sakazakii infections documented since. Year Number of cases Location References England Urmenyi and Franklin () Denmark Joker et al. () USA (Macon, GA) Monroe and Tift () USA (Indianapolis, IN) Kleiman et al. () - Netherlands Muytjens et al. () Greece Arseni et al. () USA (Boston, MA) Willis and Robinson () - Iceland Bierling et al. () USA (Memphis, TN) Simmons et al. () USA (Baltimore, MD) Noriega et al. () USA (Cincinnati, OH) Gallagher and Ball () Israel Block et al. () Israel Block et al. () - USA (Boston, MA) Lai () Israel Block et al. () Israel Block et al. () Belgium Van Acker et al. () - Israel Bar-Oz et al. () USA (Knoxville, TN) Himelright et al. () (Urmenyi and Franklin, ). In The Netherlands, eight cases of neonatal meningitis caused by E. sakazakii were reported between and (Muytens et al., ). Two of the infected infants had both necrotizing enterocolitis and meningitis and the mortality rate was %, although antibiotics were given. Three cases of E. sakazakii infections were reported between in and in Iceland (Biering et al., ). One of the three infants had Down s syndrome.

20 All three infants were fed reconstituted powdered infant formula. One died and the two recovered, but with severe neurologic sequelae. In, cases of necrotizing enterocolitis were reported in Belgium (van Acker et al., ). Fifty neonates were admitted to neonatal intensive care units in June and July,. All twelve infants with necrotizing colitis had low birth weights (< kg) and had been fed powdered infant formula. Of the neonates, ten were fed with the same formula and six presented blood, anal swab, or stomach aspirate cultures positive for E. sakazakii. E. sakazakii was isolated not only from infant formulas that were fed to the patients but also from unopened cans of products. Partial strain similarity between powdered infant formula and patient cultures was confirmed with molecular typing (AP-PCR). In the U. S., outbreaks of E. sakazakii infections were reported in Tennessee in (Simmons et al., ). The outbreak involved four infants. One had bloody diarrhea, one had sepsis, and two had both diarrhea and sepsis. In, a premature infant was infected with E. sakazakii in Knoxville, TN (CDC, a; Himelright et al., ). Forty-nine infants were examined for E. sakazakii; stool and urine samples of ten infants were positive for the organism. Powdered infant formula was suspected to be the source of E. sakazakii that caused infections. Resistance to stress Observations on the occasional presence of E. sakazakii in powdered infant formula, which is subjected to heat treatment during processing, have raised interest in determining its heat resistance. Thermal resistance has been reported to vary significantly, depending on the strain and suspending medium. Nazarowec-White and Farber (b) determined D values of E. sakazakii in reconstituted powdered infant formula. D values at,,,, and C were

21 .,.,.,., and. min, respectively. Iversen et al. () reported D values of E. sakazakii at,,,, and C to be. -.,. -.,. -.,. -., and. -. min, respectively. D values at C of strain ATCC and a clinical strain in rehydrated powdered infant formula were. and. min, respectively (Buchanan and Edelson, ). E. sakazakii in disodium hydrogen phosphate potassium buffer (ph.) have D values ranging from. to. min at C (Breeuwer et al., ). D values were definitely lower than those of cells in reconstituted powdered infant formula. Electromagnetic radiation (, MHz) was tested for effectiveness in killing E. sakazakii (Kindle et al., ). Five infant formulas were inoculated with E. sakazakii at ca. log CFU/ml and heated by microwave treatment until they boiled. E. sakazakii survived in one of the four formulas. Resistance of E. sakazakii to low water activity has been studied. Caubilla-Barron et al. () found that E. sakazakii can survive in powdered infant formula for at least months. E. sakazakii in stationary growth phase was observed to be more resistant to osmotic and desiccation stresses than were Enterobacter agglomerans, Escherichia coli, Salmonella Senftenberg, S. Typhimurium, and S. Enteritidis (Breeuwer et al., ). Stationary phase cells accumulated tremendous amounts of trehalose, compared to the cells in exponential phase, when the organism was dried at C. This may explain why E. sakazakii in stationary phase is more resistant to desiccation stress, compared to other Enterobacteriaceae, since trehalose has been determined to stabilize phospholipids in membranes, thereby protecting cells during desiccation (Crowe et al., ; Strom and Kaasen, ; Potts, ; Kempf and Bremer, ; Welsh and Herbert, ).

22 The resistance of E. sakazakii to acid ph has been studied. Ten of twelve test strains of E. sakazakii showed less than a -log reduction in tryptic soy broth h after the ph was adjusted to. with HCl; a. to >. log reduction occurred at ph. (Edelson-Mammel and Buchanan, ). E. sakazakii inoculated into milk was reported to ferment milk rapidly and reduce the ph from. to. at C in < h (Skladal et al., ). Foods from which E. sakazakii has been isolated Powdered infant formula and milk powder have been implicated as sources of E. sakazakii in outbreaks of infections (Muytjens et al., ; Biering et al., ; Simmons et al., ; Noriega et al., ; Van Acker et al., ; Bar-Oz et al., ; Block et al., ; Himelright et al., ). The first outbreak of E. sakazakii infection confirmed to be caused by powdered infant formula occurred in (CDC, a; Himelright et al., ; Weir, ). The organism was isolated from an unopened can of the formula. E. sakazakii was detected in of (. %) powdered infant formulas originating from countries (Muytjens et al., ) and in of (.%) infant formulas manufactured in South Africa, South Korea, Holland, Spain, Switzerland, USA, Belgium, Ireland, Slovenia, and UK (Iversen and Forsythe, ). Iversen and Forsythe () surveyed other infant foods and milk-based products, including dried infant foods, milk powders, lactose powders, and cheese products for presence of E. sakazakii. They found the bacterium in of (.%) dried infant foods, of (.%) milk powders, and of cheese products. Although outbreaks of E. sakazakii infections have been linked only to powdered infant formula, the organism has been isolated from various types of ready-to-eat foods, including lettuce (Soriano et al, ), four types of vegetables, cheese, minced beef, and sausage meat

23 (Leclercq et al., ), rice seed (Cottyn et al., ), beer mugs (Schindler and Metz, ), cured meat (Watanabe and Esaki, ), tofu (No et al., ), sour tea (Tamura et al., ), fermented bread (Gassem, ), tomatoes (Jung and Park, ), and alfalfa sprouts (Cruz et al., ). FOODBORNE DISEASES ASSOCIATED WITH FRESH PRODUCE In the past two decades, concomitant with an increased per capita consumption of fresh produce in the U. S., the number and frequency of outbreaks of illness associated with fresh produce have increased. Approximately % of cases of foodborne illness in the U. S. have been associated with consumption of fresh fruits and vegetables (Tauxe, ). Several genera have been involved. Examples of pathogens causing infections include E. coli O:H linked to the consumption of lettuce (Hilborn et al., ), alfalfa (Breuer et al., ), and apple juice (CDC, ), Salmonella linked to tomatoes (Cummings et al., ) and cantaloupe (CDC, b), Shigella linked to parsley (CDC, ), Vibrio cholerae on raw vegetables (CDC, ), and Cyclospora cayetanesis linked to raspberries (Herwaldt et al., ). Listeria monocytogenes, Campylobacter, Aeromonas, and Clostridium botulinum have also been concerns of fresh fruits and vegetables (Beuchat, ; Harris et al., ). Many factors have contributed to the increase fresh produce-related outbreaks. These include changes in production, handling practices, and consumption patterns (Beuchat and Ryu, ). Consequently, various sanitizing methods have been developed and applied by the produce industry to improve safety of fresh produce for consumers (Beuchat, ). The presence of E. sakazakii on produce raises concern about safety risks to immunocompromised adults. To date, however, outbreaks of E. sakazakii infections have not

24 been linked to consumption of produce. The efficacy of sanitizers routinely used in the produce industry in killing E. sakazakii on produce has not been reported. BIOFILM Definitions and mechanisms The term biofilm has been described as a biological matrix of microbial cells and extracellular substances (Bakke et al., ). Biofilms have also been defined as sessile communities of bacterial cells attached to a surface or to each other, usually embedded in polymeric substances produced by the bacteria (Marshall, ; Costerton et al., ). Attachment of bacterial cells to surfaces is followed by growth, production of exopolysaccharide, and biofilm formation (Kumar and Anand, ). Several mechanisms of biofilm formation have been proposed by researchers. Development of biofilms was detailed by Stoodley et al. () as a five-step process: () reversible attachment of cell to surface, () irreversible attachment, () development of biofilm matrix, () maturation of biofilm, and () dispersal of biofilm. Marshall et al. () viewed the biofilm formation in two- step process consisting of reversible attachment followed by irreversible attachment. Kumar and Anand () proposed that biofilm formation involved a five-stage process: () formation of conditioning film, () attachment of bacterial cells, () development of microcolonies, () biofilm formation, and () dispersion of biofilms. Prior to attachment of cells, conditioning films develop by adsorbing organic or inorganic materials to the surface which changes the surface characteristics, e.g., free energy, hydrophobicity, and electrostatic charges (Dickson and Koohmaraie, ). In food processing plants, food residues remaining on equipment surfaces may behave as conditioning films. After

25 the conditioning film forms, reversible attachment of cells occurs followed by irreversible attachment. Van der Waals forces, electrostatic interaction, and hydrophobic interaction play important roles in a reversible attachment (Chmielewski and Frank, ). Dipole-dipole interaction, hydrogen bonding, and ionic covalent bonding are involved in irreversible attachment (Bower et al., ; Briandet et al., ). During the attachment process, cells utilize nutrients from the fluid environment and form microcolonies on surfaces (Kumar and Anand, ). Microorganisms produce extracellular polymeric substances (Characklis and Marshall, ), which generally firmly attach cells to surfaces (Eginton et al., ). This leads to formation of the biofilm matrix, which is a three-dimensional structure. Once biofilms form, cells continue to grow. As the biofilms mature, however, cells begin to detach and colonize new niches (Kumar and Anand, ). The detached cells move to another location and continue the process to form biofilms (Marshall, ). Another process describing biofilm formation was proposed by Busscher and Weerkamp (). They proposed a mechanism based on distance between bacterial cells and the supporting surface. It was found that at a distance of > nm, - nm, and <. nm, Van der Waals forces only, Van der Waals forces and electrostatic interactions, and additional specific forces are involved, respectively. Factors affecting attachment and biofilm formation Factors affecting attachment of and biofilm formation by microorganisms include nutrient availability, extracellular polymeric substances, ph of the surrounding medium, and the nature of the cell surfaces (Frank, ). There are numerous studies showing that nutrient

26 availability affects attachment and biofilm formation. Hood and Zottola (b) observed that composition of growth and conditioning media influence the attachment of S. Typhimurium and L. monocytogenes to stainless steel surfaces. Wrangstadh et al. () reported that starvation of Pseudomonas sp. S cells adversely affect production of exopolysaccharide resulting in enhanced attachment of the cells. Increased nutrient levels enhance biofilm formation by L. monocytogenes (Jeong and Frank, ). Extracellular polymeric substances can change hydrophobicity of surfaces to enhance or inhibit adhesion of cells. Serratia marcescens, for example, has been found to secrete lipopeptide that can make hydrophilic surfaces hydrophobic (Matsuyama et al., ). Ryu et al. () observed that nutrient availability and exopolysaccharide produced by cells affect the ability of E. coli O:H to form biofilm on stainless steel. Properties of cells that affect attachment include cell surface structures, hydrophobicity, and surface charges. Fimbriae, the outer membrane, and S layer can play important roles in the attachment of cells (Frank, ). Hydrophobic cells have better ability than hydrophilic cells to attach to surfaces (van Loosdrecht et al., ). Cells tend to attach to surfaces that have opposite charges from themselves (Frank, ). Staphylococcus aureus, which is negatively charged, has greater ability to adhere to positively charged surfaces than to negatively charged surfaces (Hogt et al., ). However, these interactions are not always a prediction of cell attachment. Significance of bacterial attachment and biofilms in food processing and preparation areas Bacterial biofilms are known to be formed on food contact surfaces, which may result in cross-contamination of products (Zottola and Sashara, ; Wirtanen et al., ; Hood and

27 Zottola, a; Kumar and Anand, ; Frank, ). Stainless steel is a commonly used material for food contact surfaces in processing and preparation area. Numerous bacteria, including E. coli O:H (Ryu et al., ; Ryu and Beuchat, b), B. cereus (Ryu and Beuchat, a), L. monocytogenes (Hassan et al., ; Folsom et al., ), E. sakazakii (Iversen et al., ), and Pseudomonas putida (Antoniou and Frank, ) have been shown to attach and form biofilms on stainless steel surfaces. Attachment of microorganisms followed by biofilm formation on biotic or abiotic surfaces is known to enhance the resistance of cells to environmental stresses and provide protection against sanitizers (Kumar and Anand, ; Norwood and Gilmour, ; Frank et al., ; Ryu and Beuchat, b). Several studies have shown that attached microorganisms have higher resistance than planktonic cells to environmental stresses and antimicrobials. For instance, cells of L. monocytogenes on glass or stainless steel are more resistant to benzalkonium chloride (BAC), anionic acid sanitizers, and heat ( C and C), compared to planktonic cells (Frank and Koffi, ; Mafu et al., ). Delissalde and Amábile-Cuevas () reported that P. aeruginosa in biofilm showed higher resistance than non-biofilm formers to several antibiotics. Peracetic acid, mercuric chloride, and formaldehyde have been shown to be ineffective in killing microorganisms in biofilms (Carpentier and Cerf, ). Ryu and Beuchat (b) reported that the resistance of E. coli O:H to chlorine significantly increased as cells formed biofilm on the surface of stainless steel. Mechanisms of the resistance of cells to environmental stresses have been described. Extracellular polymeric substances produced by microorganisms during biofilm formation behave as protective barriers from stresses (Costerton et al., ; Lewis, ; Mah and O Toole, ). Oxidizing sanitizers, including hypochlorite and hydrogen peroxide, can be

28 neutralized and become less effective upon contact with outer layers of biofilms (de Beer et al., ; Chen and Stewart, ; Xu et al., ; Mah and O Toole, ). There have been reports that P. aeruginosa in biofilms produce significantly more β-lactamase, an antibioticdegrading enzyme, than planktonic cells (Tuomanen et al., ). The composition of cell wall protein of microorganisms in biofilm can be altered in biofilm (O Toole et al., ). Biofilm formation by E. sakazakii E. sakazakii has been found to attach to and form biofilms on silicon, latex, polycarbonate, stainless steel, glass, and polyvinyl chloride (PVC) (Iversen et al., ; Lehner et al., ). In addition, E. sakazakii has been observed to produce extracellular polysaccharide (Scheepe-Leberkühne and Wagner, ; Lehner et al., ). Attachment and biofilm formation by E. sakazakii on equipment surfaces in formula preparation or feeding areas or in produce processing plants may increase the risk of infections. Presence of E. sakazakii on a spoon, brush, and blender used for infant formula preparation has been documented in a clinical setting where neonatal infections had been reported (Simmmons et al. ; Clark et al., ; Noriega et al., ; Bar-Oz et al., ). Reuse of infant feeding equipment, e.g., infusion tubes and delivery bags, after washing with water is hypothesized to increase the risk of microbial infections (Oie and Kamiya, ). Nosocomial infections caused by E. sakazakii can occur through contaminated utensils (Martin, ).

29 SANITIZERS AND DISINFECTANTS Sanitizers used for fresh produce Various sanitizers have been applied to fresh produce for the purpose of eliminating microorganisms capable of causing diseases. Chlorinated water, chlorine dioxide, and peracetic acid-based sanitizers are among the sanitizers used in the fresh fruit and vegetable industry. Chlorine In the U.S., chlorine was first used to treat drinking water in the early th century, and then used as a sanitizer in dairy processing plants and other food processing plants (Troller, ). Chlorine has several advantages as a sanitizer, including its effectiveness in killing a broad spectrum of microorganisms, low cost, and ease of handling. Disadvantages are that chlorine is corrosive, causes discoloration, is inactivated by organic materials, and can produce off flavors (Troller, ). Water containing free chlorine concentrations of - µg/ml is used to sanitize fresh produce (Beuchat et al., ; Rogers et al., ). Free chlorine is defined as HOCl (hypochlorous acid), OCl - (hypochlorite ion) or Cl (elementary chlorine) (Weidenkopf, ). Of the free forms, HOCl exhibits the most effective bactericidal activity. Consequently, antimicrobial activity of chlorinated water depends on the amount of HOCl present (Beuchat and Ryu, ). When Cl is added to water, HOCl forms by the following reaction: Cl + H O HOCl + H + + Cl Then, HOCl dissociates into H+ and OCl in water: HOCl H + + OCl

30 Modes of antimicrobial action that have been proposed include inhibiting glucose oxidation, disrupting protein synthesis, reacting with nucleic acids, purines, and pyrimidines, inhibiting oxygen uptake, forming toxic chloramines, and changing cell permeability (Troller, ; Marriott and Gravani, ). Chlorine has been shown to be effective in reducing populations of E. coli O:H (Beuchat et al., ; Beuchat, ; Park and Beuchat, ; Fett, ; Ryu and Beuchat, b), Salmonella (Zhuang et al., ; Park and Beuchat, ; Weissinger et al., ; Beuchat et al., ; Fett, ), and L. monocytogenes (Ukuku and Fett, ; Beuchat et al., ) on fresh produce. For instance, Beuchat et al. () reported that spraying with µg/ml chlorine and soaking for min resulted in significant decreases in populations of E. coli O:H and Salmonella on apples, tomatoes, and lettuce, compared untreated produce. Beuchat et al. () observed a.-log reduction in Salmonella on tomatoes sprayed with µg/ml of chlorine. Populations of E. coli O:H, Salmonella, and L. monocytogenes on tomatoes decreased by >.,., and >. log/tomato, respectively, following treatment with µg/ml of chlorinated water (Lang et al., ). Lettuce treated with chlorine at µg/ml for min showed only.,., and. log-reductions in E. coli O:H, Salmonella, and L. monocytogenes, respectively (Lang et al., ). Chlorine dioxide Aqueous chlorine dioxide has been approved as a disinfectant used in bottling plants and food processing plants (storage and handling areas) by the U. S. Environmental Protection Agency (EPA) (). The gaseous form of chlorine dioxide was approved as a sterilant for equipment and environmental surfaces in (U. S. EPA, ). Aqueous chlorine dioxide has

31 been authorized for use as a sanitizer for whole fresh produce (U. S. FDA, ). Chlorine dioxide has been increasingly used to enhance the microbiological safety on fruit and vegetable processing plants (Synan, ; Costilow et al., ; Roberts and Reymond, ) due to its bactericidal activity over a wide ph range, rapid bactericidal action (Bernarde et al, ; ; Rav-Acha, ; McGuire and Dishinger, ), and limited reaction with organic materials (Richardson et al, ; Long et al, ). Chlorine dioxide is produced by following reactions (Troller, ; Marriott and Gravani, ): NaOCl + HCl NaCl + HOCl NaClO + HCl ClO + NaCl + H O HOCl + NaClO ClO + NaCl + H O Aqueous and gaseous chlorine dioxide have been reported to kill E. coli O:H on lettuce and baby carrots (Singh et al., ) and E. coli O:H and L. monocytogenes on green peppers (Han et al., ), apples, lettuce, strawberries, and cantaloupes (Rodgers et al., ). Du et al. () reported reduction in populations ranged from. -. log CFU/site on apple treated with. - µg/ml gaseous chlorine dioxide. Chlorine dioxide gas (. µg/ml) significantly reduced populations of Salmonella, E. coli O: H, and L. monocytogenes on fresh-cut cabbage, carrot, and lettuce (Sy et al., ). Peroxyacetic acid The U. S. EPA () registered peroxyacetic acid, an oxidizing agent, as a sanitizer for food establishments, medical facilities, and dairy and cheese processing plants. Peroxyacetic acid has been used as a sanitizer for fruit processing operations (Wisniewsky et al., ).

32 Peroxyacetic acid (peracetic acid) is produce by combining of hydrogen peroxide and acetic acid as follows. H O + CH COOH CH COO-OH + H O Peroxyacetic acid has been known to disrupt osmotic function employed by lipoprotein of cytoplasmic membrane (Block, ). It is less corrosive than chlorine-based sanitizers and does not produce toxic residues, which makes it more acceptable for use in food processing plants (Dychdala, ; Marriott and Gravani, ). Peroxyacetic acid has been used to reduce microbial populations in process water (Hilgren and Salverda, ) and on apples (Wisniewsky et al., ; Rodgers et al, ), cantaloupes (Park and Beuchat, ; Rodgers et al., ), lettuce (Beuchat et al., ; Rodgers et al., ), strawberries (Rodgers et al., ), honeydew melons, and asparagus (Park and Beuchat, ). For instance, treatment with peroxyacetic acid was shown to reduce the percentage of inoculated stainless steel chips positive for E. coli O:H by - %, compared to treatment with water (Farrell et al., ), and has a bactericidal effect to P. aeruginosa and S. aureus in biofilms (Holah et al., ). Wright et al. () reported that the population of E. coli O:H on apples was reduced by. log CFU/cm by upon treatment with µg/ml peroxyacetic acid. Wisniewsky et al. () reported that a -log CFU/apple reduction of E. coli O:H occurred on whole apples treated with Tsunami ( µg/ml), a peroxyacetic acid-based sanitizer, for min. Sanitizers and disinfectants used for food-contact surfaces Disinfection is described as a process eliminating vegetative cells of potential pathogenic bacteria on inanimate objects (Rutala and Weber, ; Exner et al., ). Surface disinfection

33 is routinely done in hospitals by cleaning with a liquid chemical disinfectant since surfaces may contribute to cross-contamination or nosocomial infections. Commercial surface disinfectants have been based largely on quaternary ammonium compounds, phenolic compounds, alcohol, chlorine, and iodophor. Quaternary ammonium compounds Quaternary ammonium compounds (quats) are one of the most frequently used chemical disinfectants to control microorganisms in clinical and industrial areas (MacBain et al., ). Quats consist of four organic groups linked to nitrogen. The chemical structure of quats is: R R N R Cl or Br R The nature of the organic groups may alter biological activity of the quats (Li et al., ). Chloride and bromide are most commonly used for commercial quats compounds. The bactericidal mechanism of quats has not been fully characterized. However, quats have been reported to have different bactericidal activity than chlorine and chlorine dioxide. In addition, bactericidal activity of quats is less affected than chlorine-based sanitizers by organic materials (Marriott and Gravani, ). Quats produce a film on surfaces that is bactericidal to vegetative cells and bacteriostatic to spores (Marriott and Gravani, ). Quats kill microorganisms by interacting with lipopolysaccharides or lipids of the cell membrane followed by penetration of cells (Russell and Gould, ). When P. aeruginosa is treated with quats, the fatty acid composition of the cell membrane is altered (Guérin-Méchin et al., ). Quats have major advantages in that they are odorless, colorless, non-irritating, less affected by presence of

34 organic materials than chlorine, non-toxic, and non-corrosive; Disadvantages are that quats may foam excessively and are ineffective against gram-negative bacteria, unstable with soap or anionic detergent, form film on surfaces, and may enhance bacterial resistance (Troller, ; Anonymous, ; Russell, ). The recommended concentration of quats used to sanitize stainless steel is g/ml (Troller, ). L. monocytogenes cells showed significant reductions in populations when treated with or g/ml of benzalkonium chloride and cetylpyridinium chloride, major components of quats, compared to treatment with water (Taormina and Beuchat, ). Planktonic and attached cells of B. cereus were reduced by. log CFU/ml and. log CFU/stainless steel chip, respectively, after exposure to quats (dialkyldimethyl ammonium chloride) at µg/ml for sec (Peng et al., ). Treatment with µg/ml of quats (benzyldimethyl tetradecylammonium chloride) caused > % mortality of planktonic cells and sessile cells of L. monocytogenes (Chavant et al., ). Antibiotic-susceptible isolates of E. coli, S. aureus, S. epidermidis, and E. cloacae were not detected after exposure to quats (.% alkyl dimethyl ammonium chlorides) for min (Guimarães et al., ). Phenolic compounds Another group of compounds commonly used for disinfecting abiotic surfaces are phenols. Phenols are effective in killing bacteria, fungi, and many viruses. The structure of phenol is as follows: OH

35 Phenolic compounds are less inactivated by organic soil than is hypochlorite (Tyler and Ayliffe, ). Use of phenolic compounds to disinfect floors has been shown to decrease bacterial populations (Vesley and Michaelsen, ). Application of phenolic germicides to hospital floors reduced microbial populations in the environment for months (Kundsin and Walter, ). Phenolic disinfectant was less effective than quats in inactivating isolates of gramnegative bacteria from a hospital (Navajas et al., ). Phenolic disinfectants retain more activity in the presence of organic material than do iodine and chlorine-containing disinfectants (Bloomfield and Miller). Gram-negative bacteria, including members of the Enterobacteriaceae family isolated from hospitals, have been reported to be resistant to quats and phenolic disinfectants (Russell et al., ; Hammond et al., ). Hydrogen peroxide Hydrogen peroxide (H O ) has been used as an antiseptic. However, it rapidly decomposes to water and oxygen upon contact with organic materials, resulting in no residual bactericidal effects (Lück and Jager, ). Hydrogen peroxide is declared to be Generally Recognized as Safe (GRAS) ( CFR.) (U. S. FDA, ) and allowed to be used for packaging and surface sterilization in food processing plants ( CFR.) (U. S. FDA, ). Hydrogen peroxide does not behave as an oxidizer but produces powerful oxidants, viz., hydroxyl radical, singlet oxygen, and superoxide radicals (Davidson and Branen, ). Of the these oxidants, the hydroxyl radical (HO) plays an important role in toxicity to bacteria (Imlay and Linn, ). It increases lipid peroxidation and ion permeability of the cell membrane (Anzai et al., ).

36 Hydrogen peroxide has been shown to have bactericidal effects on various foodborne pathogens. Populations of planktonic cells of L. monocytogenes were reduced by. and. log CFU/ml after exposure to a.% hydrogen peroxide solution for and min, respectively (Robbins et al., ). These researchers also showed a.-log reduction of L. monocytogenes in biofilms on stainless steel chips treated with a % solution of hydrogen peroxide. S. Stanley freshly attached to the surface of cantaloupe showed ca. -log CFU/cm decrease in population following treatment with a % solution of hydrogen peroxide (Ukuku and Sapers, ). OBJECTIVES The objectives of the research reported in this dissertation are as follows:. Study survival characteristics of E. sakazakii in fresh-cut produce and unpasteurized juice as affected by storage temperature Outbreaks of E. sakazakii infections have been associated with infant formulas but documented presence of the organism in a wide range of ready-to-eat foods, including lettuce, tomatoes, bean sprouts, and other raw vegetables raises interest in knowing more about its behavior on produce. Although E. sakazakii has been reported to present on vegetables, thereby raising a potential public health concern, to date, outbreaks of infection linked to fresh produce have not been documented. Factors influencing survival and growth of E. sakazakii in produce have not been studied. Findings from this study will provide information on the level of potential risk of E. sakazakii infections that may be associated with fresh-cut produce and produce juice.

37 . Determine survival characteristics of E. sakazakii on the surface of whole produce as affected by storage temperature and its resistance to produce sanitizers, including chlorine, chlorine dioxide, and a peracetic acid-based sanitizer E. sakazakii has caused infections in neonates, infants, and elderly immunocompromised adults. Sources of the pathogen have most frequently implicated powdered infant formulas. Observations from this study will provide insights to predicting E. sakazakii survival on whole (uncut) produce and the efficacy of sanitizers in killing the bacterium, thereby enabling the development of effective treatments to reduce the risk of E. sakazakii infections in immunocompromised individuals.. Investigate influences of temperature and nutrient availability on attachment and biofilm formation by E. sakazakii on surfaces of stainless steel and enteral feeding tubes E. sakazakii infections in neonates have been associated with consumption of reconstituted infant formula but the pathogen also has been isolated from a wide range of foods, including meat, dairy, cereal, and vegetable products. The pathogen has been reported to form biofilms on stainless steel, latex, polycarbonate, and silicon surfaces, but environmental conditions affecting attachment and biofilm formation on abiotic surfaces such as stainless steel and enteral feeding tubes have not been described. This study will provide insights to the attachment of and biofilm formation by E. sakazakii on stainless steel and enteral feeding tubes upon exposure to various nutrients and at various temperatures.